Can 3D artificial intelligence models outshine 2D ones in the detection of intracranial metastatic tumors on magnetic resonance images?

BACKGROUND: This study aimed to compare the prediction performance of two-dimensional (2D) and three-dimensional (3D) semantic segmentation models for intracranial metastatic tumors with a volume ≥ 0.3 mL. METHODS: We used postcontrast T1 whole-brain magnetic resonance (MR), which was collected from Taipei Veterans General Hospital (TVGH). Also, the study was approved by the institutional review board (IRB) of TVGH. The 2D image segmentation model does not fully use the spatial information between neighboring slices, whereas the 3D segmentation model does. We treated the U-Net as the basic model for 2D and 3D architectures. RESULTS: For the prediction of intracranial metastatic tumors, the area under the curve (AUC) of the 3D model was 87.6% and that of the 2D model was 81.5%. CONCLUSION: Building a semantic segmentation model based on 3D deep convolutional neural networks might be crucial to achieve a high detection rate in clinical applications for intracranial metastatic tumors.

Expression and characterization of the biofilm-related and carnosine-hydrolyzing aminoacylhistidine dipeptidase from Vibrio alginolyticus.

The biofilm-related and carnosine-hydrolyzing aminoacylhistidine dipeptidase (pepD) gene from Vibrio alginolyticus was cloned and sequenced. The recombinant PepD protein was produced and biochemically characterized and the putative active-site residues responsible for metal binding and catalysis were identified. The recombinant enzyme, which was identified as a homodimeric dipeptidase in solution, exhibited broad substrate specificity for Xaa-His and His-Xaa dipeptides, with the highest activity for the His-His dipeptide. Sequence and structural homologies suggest that the enzyme is a member of the metal-dependent metallopeptidase family. Indeed, the purified enzyme contains two zinc ions per monomer. Reconstitution of His.Tag-cleaved native apo-PepD with various metal ions indicated that enzymatic activity could be optimally restored when Zn2+ was replaced with other divalent metal ions, including Mn2+, Co2+, Ni2+, Cu2+ and Cd2+, and partially restored when Zn2+ was replaced with Mg2+. Structural homology modeling of PepD also revealed a 'catalytic domain' and a 'lid domain' similar to those of the Lactobacillus delbrueckii PepV protein. Mutational analysis of the putative active-site residues supported the involvement of His80, Asp119, Glu150, Asp173 and His461 in metal binding and Asp82 and Glu149 in catalysis. In addition, individual substitution of Glu149 and Glu150 with aspartic acid resulted in the partial retention of enzymatic activity, indicating a functional role for these residues on the catalysis and zinc ions, respectively. These effects may be necessary either for the activation of the catalytic water molecule or for the stabilization of the substrate-enzyme tetrahedral intermediate. Taken together, these results may facilitate the design of PepD inhibitors for application in antimicrobial treatment and antibody-directed enzyme prodrug therapy.